Food Research International 41 (2008) 973–979

Contents lists available at ScienceDirect

Food Research International

journal homepage: www.elsevier .com/ locate / foodres

Acute ingestion of yerba mate infusion (Ilex paraguariensis) inhibits plasmaand lipoprotein oxidation

Edson L. da Silva a,*, Terezinha J.C. Neiva a, Mutsuko Shirai b, Junji Terao c, Dulcinéia S.P. Abdalla d

a Department of Clinical Analyses, Health Sciences Center, Federal University of Santa Catarina, 88040970 Florianópolis, Santa Catarina, Brazilb Department of Nutritional Sciences, Faculty of Human Ecology, Yasuda Women’s University, Hiroshima, Japanc Department of Food Science, Institute for Health Biosciences, The University of Tokushima Graduate School, Japand Department of Clinical and Toxicological Analyses, School of Pharmaceutical Sciences, University of Sao Paulo, Brazil

a r t i c l e i n f o

Article history:Received 29 July 2008Accepted 9 August 2008

Keywords:Ilex paraguariensisMatePhenolsLipid peroxidationPlatelet aggregationHumans

0963-9969/$ - see front matter � 2008 Elsevier Ltd. Adoi:10.1016/j.foodres.2008.08.004

* Corresponding author. Tel.: +55 48 37218053; faxE-mail addresses: edson@ccs.ufsc.br, dasilvael@hot

a b s t r a c t

Yerba mate extract (Ilex paraguariensis) is a source of phenolic compounds that possesses in vitro antiox-idant activities and may contribute to a reduction in the risk of cardiovascular disease. In this study weexamined the acute effects of the consumption of mate infusion on ex vivo plasma and low-density lipo-protein (LDL) oxidation, plasma antioxidant capacity, and platelet aggregation. Twelve healthy fastedsubjects ingested 500 mL of mate infusion and blood samples were collected before and 1 h after mateintake. Lipid peroxidation of plasma and LDL was monitored by the measurement of cholesteryl-esterhydroperoxides (CE-OOH) and cholesterol oxides. The plasma antioxidant capacity was measured as fer-ric-reducing antioxidant potential (FRAP). Platelet aggregation was evaluated in platelet-rich plasmastimulated with adenosine diphosphate and coagulation was tested in platelet-poor plasma. Ingestionof mate infusion diminished the ex vivo oxidizability of both plasma and LDL particles. After mate intake,the CE-OOH levels were around 50% lower in plasma oxidized with copper or 2,20-azobis[2-amidine-pro-pane-hydrochloride] (AAPH) and the lag time to plasma oxidation increased 2-fold (P < 0.05). Copper-and AAPH-induced LDL peroxidation were also inhibited by around 50% and 20%, respectively, after mateconsumption (P < 0.05). The levels of various oxysterols were significantly reduced in oxidized-plasmaand LDL (P < 0.05) and FRAP increased by 7.7% after mate intake (P < 0.01). However, mate consumptiondid not inhibit platelet aggregation or blood coagulation. In summary, intake of yerba mate infusionimproved the antioxidant capacity and the resistance of plasma and LDL particles to ex vivo lipidperoxidation.

� 2008 Elsevier Ltd. All rights reserved.

1. Introduction

The results of epidemiological studies over the past few yearshave shown the beneficial effects of an increased consumption ofvegetables and fruits, which are enriched in vitamins E and C(Giuliano, 2000; Rimm & Stampfer, 1997), flavonoids and otherphenolic compounds (Geleijnse, Launer, Kuip, Hotiman, & Witt-eman, 2002; Huxley & Neil, 2003), in reducing risk of developingcoronary heart disease. The development of atherosclerosis isthought to be critically dependent on LDL oxidation (Stocker & Kea-ney Jr., 2004) and platelet activation and aggregation (Aviram,1992). In this context, it has been suggested that the inverse asso-ciation between flavonoid intake and atherosclerosis shown insome epidemiological studies might be explained, in part, by theantioxidant ability of phenolic compounds (Geleijnse et al., 2002;Huxley & Neil, 2003).

ll rights reserved.

: +55 48 37219542.mail.com (E.L. da Silva).

Ilex paraguariensis (St. Hil., Aquifoliaceae) infusion is a potentialsource of polyphenols. Aqueous extract of I. paraguariensis, knownas mate or yerba mate, is a typical antioxidant-containing beveragelargely consumed in several South American countries, includingsouthern Brazil, Uruguay, Paraguay and Argentine. The mate ex-tract, which is prepared as a hot infusion of the dried and mincedleaves of I. paraguariensis, is very prized for its bitter taste andstimulating properties. Moreover, yerba mate presents a healthynutrition profile. It has been claimed that it is difficult to find aplant anywhere in the world similar to mate in nutritional valueand that yerba mate contains practically all of the vitamins andminerals necessary to sustain life. For example, yerba mate pos-sesses the vitamins A, B1, B2, niacin (B3), B complex, C, and E,the minerals potassium, magnesium, calcium, manganese, iron,selenium, phosphorus, and zinc, several amino acids, 196 volatilechemical compounds, of which 144 are also found in green tea,and various polyphenols, mainly caffeoyl derivatives (such aschlorogenic acid), and some flavonoids (rutin, quercetin, kaempfer-ol, and luteolin) (Bastos, Oliveira, Matsumoto, Carvalho, & Ribeiro,

974 E.L. da Silva et al. / Food Research International 41 (2008) 973–979

2007; Heck & de Mejia, 2007). In addition to polyphenols, yerbamate infusion contains saponins, caffeine and theobromine (Bastoset al., 2007; Heck & de Mejia, 2007). In folk medicine, mate infusionhas been used for the treatment of arthritis, rheumatism and otherinflammatory diseases, headache, obesity, hypertension, andhepatic and digestive disorders. Recent findings of our laboratoryhave shown that an aqueous extract of I. paraguariensis can inhibitthe progression of atherosclerosis in cholesterol-fed rabbits(Mosimann, Wilhelm-Filho, & Silva, 2006). Additionally, we havereported that consumption of mate extract reversed the endothe-lial dysfunction in atherosclerotic aorta of LDL receptor-deficientmice (Felippi, Wilhelm-Filho, Ribeiro-Do-Valle, & Silva, 2006). Fur-thermore, it has been demonstrated that I. paraguariensis promotesin vitro vascular relaxation (Muccillo-Baisch, Johnston, &Paganini-Stein, 1998), exhibits superoxide scavenging ability andinhibits lipid peroxidation in several in vitro systems, includingLDL oxidation (Campos, Escobar, & Lissi, 1996; Filip, Lotito, Ferraro,& Fraga, 2000; Gugliucci & Stahl, 1995). However, studies on theantioxidant ability of I. paraguariensis in humans are scarce. Gug-liucci (1996) showed for the first time that ingestion of I. paraguari-ensis extract inhibited ex vivo plasma oxidation of healthyvolunteers. Oxidation of LDL particles was also inhibited in wholeplasma but not after LDL isolation (Gugliucci, 1996). Recently,Menini, Heck, de Mejia, and Gugliucci (2007) also demonstratedan increased activity of the antioxidant enzyme paraoxonase-1 inthe plasma of healthy subjects after acute intake of I. paraguariensisextract. Here, we confirmed the results of ex vivo plasma lipid per-oxidation inhibition following mate intake and extend them to iso-lated LDL. These findings suggest that antioxidants from yerbamate were absorbed and might remain adhered to the LDL parti-cles. In addition, enhancement of the antioxidant capacity of plas-ma was also shown. Overall, our data indicate that besides itsstimulant and nutritional proprieties yerba mate might be consid-ered an important source of antioxidants to humans.

2. Material and methods

2.1. Plant material and chemicals

Yerba mate infusion was prepared from commercially availableloose leaves of I. paraguariensis, purchased from Leao Junior SA(Curitiba-PR, Brazil). A specimen of the plant was identified asauthentic I. paraguariensis in the Botanical Department of the Fed-eral University of Santa Catarina. Caffeic, chlorogenic, ferulic, andvanillic acids, 3,5-di-tert-butyl-4-hydroxytoluene (BHT), Folin re-agent, N,O-bis-trimethylsilyltrifluoroacetamide (BSTFA), Troloxand albumin were purchased from Sigma–Aldrich (St. Louis, MO,USA). 2,20-Azobis[2-amidine-propane-hydrochloride] (AAPH) wassupplied by Eisai Pharmaceutical Co. (Tokyo, Japan). Adenosinediphosphate (ADP) was obtained from Chrono-log Corp. (Haver-town, PA, USA). Thromboplastin and ellagic acid/phospholipidwere purchased from Organon Technika (Durham, NC, USA). Sol-vents of high-performance liquid chromatography (HPLC) gradeand 2,4,6-tri(2-piridil)-s-triazine (TPTZ) were supplied by Fluka(Milwaukee, WI, USA). Other chemicals used were of the purestanalytical grade available.

2.2. Preparation of yerba mate infusion

Infusion of yerba mate was prepared by mixing boiling waterand dried and minced leaves of commercial yerba mate in the pro-portion of 50 mg/mL. After 10 min extraction, the mixture was fil-tered and drunk immediately by the volunteers. The concentrationof solid substances in the aqueous extract of yerba mate was deter-mined by weighing 1 mL of extract after complete evaporation.

Powdered extract was obtained by speed-vacuum centrifugation(Eppendorf, Hamburg, Germany) of the aqueous mate extract andused in the in vitro studies.

2.3. Determination of total phenol content in yerba mate infusion

The level of total phenols in the infusion of yerba mate wasdetermined spectrophotometrically using Folin–Ciocalteau reagent(Singleton, Orthofer, & Lamuela-Raventos, 1999) with chlorogenicacid as a standard. Five infusions, in triplicate (CV < 3%), were usedfor the measurement of total phenols.

2.4. Subjects and experimental design

Twelve healthy volunteers (6 male and 6 female), with a meanage of 29.5 years (range of 21–38 years), were recruited from thegeneral student and staff population of the research laboratories.Exclusion criteria of the potential volunteers were the use of anymedication, alcohol or antioxidant supplements, a history of majorillness, including bleeding or thrombosis, heart disease, liver andrenal diseases, diabetes mellitus, and stomach or intestinal disor-ders, and a body mass index > 25 kg/m2. The Federal Universityof Santa Catarina and the University of São Paulo Ethics Commit-tees approved the study and all participants gave written, informedconsent.

In order to verify the acute effects of yerba mate intake on theinhibition of ex vivo lipid peroxidation and platelet aggregation,the subjects were instructed not to drink black or green tea or cof-fee during the 2 days immediately prior the study. The volunteerswere 12 h fasted and drank 500 mL of yerba mate infusion. A bloodsample was drawn from the antecubital vein by venipuncture intoheparin and citrate tubes just before the subject started drinkingthe mate infusion (to determine baseline levels) and 1 h after fin-ishing the drink. Heparinized plasma was isolated by centrifuga-tion at 4 �C, 750g, 15 min, and used immediately in theexperiments related to oxidation, antioxidant capacity and forLDL isolation. Citrated blood was used for platelet isolation andcoagulation assays.

2.5. LDL and plasma lipid peroxidation

LDL was isolated from heparinized plasma by discontinuousdensity-gradient ultracentrifugation as described previously bySilva, Tsushida, and Terao (1998). Protein concentration in theLDL solution was determined by the method of Bradford (1976),using bovine serum albumin as the standard. The purity of theLDL preparation was verified by SDS–polyacrylamide gel electro-phoresis. Isolated LDL was used immediately in the experiments.For ex vivo oxidation assays, the whole plasma and isolated LDL(0.2 mg protein/mL) were incubated with 500 and 5 lmol/L CuCl2,respectively, or 50 and 4 mmol/L AAPH, respectively, at 37 �C withcontinuous shaking for up to 6 h in the dark. Aliquots of the mix-ture were taken at the different times given in the figure legendand oxidation was stopped by addition of 40 lmol/L BHT and5 mmol/L EDTA (final concentrations). The oxidation progresswas monitored by the measurement of the production of choleste-ryl-ester hydroperoxide (CE-OOH) using high-performance liquidchromatography (HPLC). Lag time was measured from the plot ofCE-OOH against time and was defined as the intercept betweenthe baseline and the tangent of the CE-OOH curve during thepropagation phase. In addition, the oxidation progress was alsomonitored by the measurement of cholesterol oxides by gaschromatography. For this purpose, plasma and LDL particles wereincubated at 37 �C for up to 12 and 6 h, respectively, with copperor AAPH.

E.L. da Silva et al. / Food Research International 41 (2008) 973–979 975

2.6. Determination of CE-OOH in plasma and LDL

To determine the concentration of CE-OOH, aliquots (100 lL)were withdrawn from the plasma or LDL incubation systems atregular time intervals as indicated and mixed with 2 mL acetoni-trile containing 0.02% BHT in amber tubes. The mixture was vor-tex-mixed for 2 min, 4 mL of Chelex-treated hexane were thenadded, the tubes were agitated in a vortex for 2 min and centri-fuged at 1000g for 10 min. The hexane phase was collected andevaporated by nitrogen gas. The residue was dissolved with meth-anol:butanol 2:1 (v/v), filtered through a 0.22 lm filter (Millipore,São Paulo, Brazil) and aliquots were subject to CE-OOH analysiswith reversed-phase HPLC (Perkin–Elmer Instruments, Bucking-hamshire, England), using a LC18DB column (25 cm � 4.6 mm �5 lm; Supelco, Bellefonte, PA, USA) with a chemiluminescencedetector. The eluting solvent was composed of methanol:butanol2:1 (v/v) and was used at a flow rate of 1.0 mL/min. The derivatiza-tion solution for the post-column reaction was a 1:1 mixture ofmethanol and borax buffer pH 10.0, containing 1.5 mg/L of micro-peroxidase and 177.2 mg/L of isoluminol, at a rate of 1 mL/min. Theconcentration of CE-OOH was calculated from the standard curveof the hydroperoxy-derivative of cholesteryl linoleate. All determi-nations were carried out in duplicate.

2.7. Determination of cholesterol oxides in plasma and LDL

Cholesterol oxides were determined in the lipid fraction of theplasma or LDL which was extracted with chloroform–methanol(1:2 v/v), according to Ferderbar et al. (2007). This methodologydoes not separate 5-cholesten-3b,25-diol (25-OH) from 5-chole-sten-3b-ol-7-one (7-keto). All determinations were carried out induplicate.

2.8. Plasma antioxidant capacity

The antioxidant capacity of the plasma was determined usingthe ferric-reducing potential (FRAP) assay as previously describedby Benzie and Strain (1996). The antioxidants present in the plas-ma obtained before and 1 h after mate intake were evaluated asreducers of Fe3+ to Fe2+, which is chelated by 2,4,6-tri(2-pyri-dyl)-s-triazine (TPTZ) to form the Fe2+-TPTZ complex withmaximum absorbance at 593 nm. The results were calculatedusing a standard curve prepared with different concentrationsof Trolox, a vitamin E water-soluble analogue, and were ex-pressed as Trolox equivalents. All determinations were carriedout in duplicate.

2.9. Platelet isolation and aggregation

For platelet studies, venous blood was collected from volunteersbefore and after yerba mate intake, into 3.8% sodium citrate at a ra-tio of 9:1 v/v. Platelet-rich plasma was prepared by low-speed cen-trifugation (400g for 7 min) at 22 �C, and the remaining samplewas re-centrifuged at 900g for 10 min to obtain platelet-poor plas-ma for coagulation tests. Adenosine diphosphate (ADP) was usedas the aggregating agent at a concentration of 6 lmol/L (aggrega-tion amplitude of up to 60%). Platelet aggregation was determinedin platelet-rich plasma by the turbidimetric method, as previouslydescribed (Neiva, Morais, Polack, Simões & D’amico, 1999), in aChrono-log 660VS aggregometer (Havertown, PA, USA). Aliquotsof platelet suspension (400 lL) were transferred into a small cuv-ette and stirred at a constant speed of 180g at 37 �C. The extentof aggregation was recorded continually for 5 min after additionof the agonist and the results are expressed as the percentage ofplatelet aggregation.

2.10. Measurement of blood coagulation

Citrated blood, collected before and 1 h after yerba mate intake,was used for coagulation tests in platelet-poor plasma. Prothrom-bin time (PT) and activated partial thromboplastin time (APTT)were evaluated according to the method described by Triplett,Harms, and Koepke (1978), using thromboplastin and ellagicacid/phospholipid reagents (Organon Technika, Durham, NC,USA), and a coagulation system (Net Lab Digital, São Paulo, SP,Brazil).

2.11. In vitro inhibition of human plasma and LDL lipid peroxidationby mate extract and pure phenols

To evaluate the antioxidant activities of potential metabolites ofchlorogenic acid, a major phenolic compound in yerba mate,in vitro assays were performed using caffeic, ferulic and vanillicacids. Blood was collected from three to five healthy fasted donorsin heparinized tubes. Pooled plasma, obtained after centrifugationof blood (1500g, for 15 min), was 4-fold diluted with phosphatebuffer saline, pH 7.4. LDL particles were isolated from the plasmaby ultracentrifugation as previously described (Silva et al., 1998).Oxidation of pooled plasma and LDL (0.2 mg protein/mL) was car-ried out at 37 �C for 3 h in the dark with continuous shaking with200 or 5 lmol/L CuCl2 for plasma and LDL, respectively, in the ab-sence (control) or presence of increasing amounts (0.5–30.0 lg/mL) of yerba mate aqueous extract (eq. phenols) or chlorogenic,caffeic, ferulic, or vanillic acids. The oxidation progress was moni-tored by the measurement of CE-OOH.

2.12. In vitro effects of yerba mate extract on platelet aggregation andblood coagulation

To confirm the absence of yerba mate effects on the plateletaggregation and blood coagulation after intake of mate, in vitrostudies were performed. Platelet-rich and platelet-poor plasmawere obtained from citrated blood of healthy control subjects. Priorto aggregation and coagulation assays, the platelet-rich plasmacontaining 2.5 � 108 cells/mL or 100 lL of platelet-poor plasmawere incubated at 37 �C for 5 min with 800 lg/mL mate extractor saline solution (control). Afterward, the platelet aggregationand coagulation tests were analyzed as described above.

2.13. Statistical analysis

Results are expressed as means ± SEM. Baseline and post-treat-ment parameters were compared using the paired Student t-test(Sokal & Rohlf, 1995). For the in vitro studies, data were analyzedusing the unpaired Student t-test. The minimum sample sizeneeded to detect a statistically significant difference (alpha < 0.05)was calculated based on the power of 0.8. To detect a 10% differ-ence in plasma and LDL lipid peroxidation the sample size requiredwas estimated to be 12 individuals, assuming a standard deviationof 11 nmol/mg protein or 11 nmol/L for lipid peroxidation response(Sokal & Rohlf, 1995). All statistical analyses were performed usingSigmaStat Version 3.2 (SPSS Inc., Chicago, Ill.). Differences ofP < 0.05 were considered significant.

3. Results

3.1. Characterization of yerba mate aqueous extract

The mate extract prepared by the addition of hot water overdried and minced leaves of commercial I. paraguariensis in the pro-portion of 50 mg/mL yielded 16.3 ± 0.27 mg/mL of solid sub-

Table 1Plasma susceptibility to ex vivo lipid peroxidation and antioxidant capacity before and1 h after ingestion of yerba mate infusionA

Measurements Yerba mate intake

Before After

Lag time to CE-OOH formation (min)Copper-induced oxidation 120 ± 2.9a 240 ± 5.2b

AAPH-induced oxidation 60 ± 1.5a 180 ± 4.3b

AAPH-induced oxysterols formation (lg/mL)5-Cholestene-3b,7a-diol (7a-OH) 26.2 ± 2.3a 23.3 ± 1.9a

5-Cholestene-3b,7 a -diol (7b-OH) 36.8 ± 2.1a 33.1 ± 1.8a

7-Keto, 5-cholesten-3b-ol-7-one (7-keto) + 5-Cholesten-3b,25-diol (25-OH)B

35.2 ± 1.6a 28.5 ± 2.3b

Cholestan-5a,6a -epoxy-3b-ol (5b,6b-epoxy) 23.2 ± 1.0a 19.9 ± 1.5a

Cholestan-5b,6b-epoxy-3b-ol (5b,6b-epoxy) 9.03 ± 0.8a 6.13 ± 0.5b

Cholestan-3b,5a,6b-triol (3b,5a,6b -triol) 7.80 ± 2.0a 8.57 ± 1.3a

Antioxidant capacity (lmol/L) 499.4 ± 10.3a 537.9 ± 18.0b

A Data are expressed as mean ± SEM of measurements from 12 individuals. Val-ues in the same row that do not share the same superscript letter are significantlydifferent (paired t-test, P < 0.05). Plasma samples were oxidized with 500 lmol/Lcopper ions or 50 mmol/L AAPH for 6 h for the cholesteryl-ester hydroperoxide (CE-OOH) measurements. Cholesterol oxides were determined in plasma oxidized withAAPH for 12 h. Antioxidant status was measured based on the ferric-reducing

976 E.L. da Silva et al. / Food Research International 41 (2008) 973–979

stances. Total phenol content in the aqueous extract of mate was2.93 ± 0.12 mg/mL. Therefore, the total polyphenol intake fromthe yerba mate infusion was �1.5 g.

3.2. Effect of yerba mate infusion consumption on ex vivo plasma lipidperoxidation

Fig. 1 shows the kinetic analysis of human plasma lipid peroxi-dation, as measured by the accumulation of CE-OOH. Plasma ob-tained after consumption of mate infusion showed a significantlylower susceptibility to copper ion- or AAPH-induced lipid peroxi-dation compared with plasma obtained before mate intake. Theconsumption of mate infusion increased approx 2-fold the lag timeof copper-induced plasma lipid peroxidation (120 ± 2.9 min, beforemate vs. 240 ± 5.2 min after mate; Table 1). In addition, after 3 and6 h copper-induced plasma oxidation, the level of CE-OOH wasapproximately 60% and 41%, respectively, lower in plasma ob-tained after mate consumption than that collected before mate in-take (P < 0.05). Furthermore, the lag time required for the initiationof plasma oxidation induced by AAPH increased around 3-fold,from 60 ± 1.5 min before mate intake to 180 ± 4.3 min 1 h after in-

0 2 5 6

0

20

40

60

80

** * *

**

A

Seru

m C

E-O

OH

(nm

ol/L

)Se

rum

CE-

OO

H (n

mol

/L)

Incubation time (h)

Incubation time (h)

0

40

80

120

160

*

*

*

*

B Before mate intake After mate intake

Before mate intake After mate intake

431

0 2 5 6431

Fig. 1. Effect of acute consumption of yerba mate infusion on the susceptibility ofplasma to ex vivo lipid peroxidation induced by (A) copper ions or (B) AAPH. Plasmasamples from 12 fasted volunteers, obtained before and 1 h after ingestion of500 mL mate extract, were oxidized with 500 lmol/L CuCl2 or 50 mmol/L AAPH for6 h at 37 �C. The oxidation was monitored by the measurement of cholesteryl-esterhydroperoxide (CE-OOH). Values are mean ± SEM. * P < 0.05 compared with resultsbefore mate intake using the paired t-test.

ability of the plasma (FRAP) assay (Benzie & Strain, 1996) and results are expressedas Trolox equivalent.

B The gas chromatography methodology used does not separate 25-OH from 7-keto.

take (P < 0.05; Table 1). The CE-OOH formation was inhibited byaround 44% and 19% after 3 and 6 h oxidation with AAPH, respec-tively (P < 0.05).

The content of cholesterol oxides in ex vivo AAPH-oxidized plas-ma is shown in Table 1. In spite of the great variability of data, asmall but still significant inhibition of 7-keto-cholesterol plus 25-hydroxycholesterol and cholesterol-5b,6b-epoxide formation wasseen in oxidized plasma obtained after ingestion of mate infusion(P < 0.05). Additionally, the consumption of yerba mate inhibitedapproximately 2.7-fold the accumulation of cholesterol-5b,6b-epoxide in copper-induced plasma oxidation (before: 6.0 ± 0.6 vs.after: 2.0 ± 0.5 lg/mL, P < 0.05). However, the levels of other oxys-terols studied were not decreased by mate intake (Table 1).

3.3. Effect of yerba mate infusion consumption on ex vivo LDL lipidperoxidation

The susceptibility of LDL to copper ion- or AAPH-induced lipidperoxidation was significantly reduced after consumption of yerbamate infusion, as shown by a 52% inhibition of the CE-OOH forma-tion after 3 h incubation with copper ions (69.6 ± 15.0 nmol/mgprotein before mate intake vs. 33.4 ± 6.4 nmol/mg protein aftermate intake, P < 0.05; Fig. 2A). After 6 h incubation with copperions, a 20% inhibition of LDL oxidation following mate intake wasobserved, however, this difference did not reach statistical signifi-cance due to the high variability in the data. The consumption ofmate infusion also protected the LDL against AAPH-induced lipidperoxidation as indicated by the 20.8% decrease in CE-OOH levelsafter 3 h oxidation (P < 0.05; Fig. 2B). On the other hand, the lagtime required for the initiation of LDL oxidation by copper or AAPHwas not increased by mate infusion consumption (results notshown here).

Furthermore, the acute ingestion of yerba mate infusion inhib-ited the accumulation of oxysterols in LDL particles oxidized bycopper or AAPH for 6 h (Fig. 3). The content of 7a- and7b-hydrox-ycholesterol and cholestane-3b,5a,6b-triol decreased from approx-imately 394, 445 and 151 lg/mg LDL protein, respectively, beforemate intake to 266, 382 and 91 lg/mg LDL protein, respectively,1 h after mate consumption (P < 0.05; Fig. 3A). In addition, the lev-els of 7b-hydroxycholesterol, 25-hydroxycholesterol plus 7-keto

0

100

200

300

*

A

63Incubation time (h)

CE-

OO

H (n

mol

/mg

prot

.)

0

50

100

150

200

250

*

B

63Incubation time (h)

CE-

OO

H (n

mol

/mg

prot

.)

Before mate intake After mate intake

Before mate intake After mate intake

Fig. 2. Effect of ingestion of yerba mate infusion on ex vivo LDL lipid peroxidation.LDL particles were isolated by ultracentrifugation from the plasma of 12 fastedvolunteers before and 1 h after ingestion of 500 mL mate infusion and incubatedwith (A) 5 lmol/L CuCl2 or (B) 4 mmol/L AAPH for 3 and 6 h at 37 �C. Lipidperoxidation was monitored by the measurement of cholesteryl-ester hydroperox-ide (CE-OOH). Values are mean ± SEM. * P < 0.05 compared with before mate intakeusing the paired t-test.

0

200

400

600

*

*

*

A

25-OH/7-keto

3 ,5α ,6 -triol

5α,6α -epoxy

5 ,6 -epoxy

7 -OH7α β β ββ

-OH

Before Mate Intake After Mate Intake

0

200

400

600

*

**

B

LDL-

oxys

tero

ls (μ

g/m

g pr

ot.)

LDL-

oxys

tero

ls (μ

g/m

g pr

ot.)

Before Mate Intake After Mate Intake

α

25-OH/7-keto

3 ,5α,6 -triol

5α,6α-epoxy

5 ,6 -epoxy

7 -OH7α β β ββ

-OH α

Fig. 3. Inhibitory effect of intake of 500 mL yerba mate infusion on ex vivo LDLperoxidation as measured by the oxysterol content. LDL particles were isolatedfrom the plasma of 12 fasted volunteers before and 1 h after ingestion of mateinfusion and incubated with (A) 5 lmol/L copper or (B) 4 mmol/L AAPH at 37 �C for6 h. Values are mean ± SEM. * P < 0.05 compared with before mate intake using thepaired t-test. 7a-OH = 5-cholestene-3b,7a-diol; 7b-OH = 5-cholestene-3b,7b-diol;7-keto = 5-cholesten-3b-ol-7-one; 25-OH = 5-cholesten-3b,25-diol; 5a,6a-epoxy = cholestan-5a,6a-epoxy-3b-ol; 5b,6b-epoxy = cholestan-5b,6b-epoxy-3b-ol; 3b,5a,6b-triol = cholestan-3b,5a,6b-triol.

E.L. da Silva et al. / Food Research International 41 (2008) 973–979 977

cholesterol, and cholesterol-5b,6b-epoxide in AAPH-induced LDLoxidation were significantly lower after ingestion of mate(P < 0.05; Fig. 3B).

3.4. Plasma antioxidant capacity (FRAP assay)

Plasma antioxidant capacity, as measured by the ferric-reducingantioxidant potential (FRAP), was 499.4 ± 10.3 lmol/L Troloxequivalents before yerba mate intake. One hour after mate con-sumption, FRAP increased slightly but significantly, with a 7.7%enhancement (537.9 ± 18.0 lmol/L Trolox equivalents; P < 0.01)(Table 1).

3.5. Platelet aggregation and blood coagulation: ex vivo and in vitrostudies

In folk medicine, mate infusion has been used to improve bloodfluidity. To verify whether acute ingestion of yerba mate infusioncan inhibit platelet aggregation and blood coagulation, platelet-rich plasma aggregation and PT and APTT tests in citrated plasmawere carried out before and after mate intake. One hour after yerbamate consumption, the ADP-induced platelet aggregation was notreduced compared to the values before ingestion of yerba mate

infusion (71.3 ± 1.5% after vs. 71.6 ± 0.9% before). Additionally, pro-thrombin time (12.7 ± 0.04% after vs. 12.0 ± 0.08% before) and acti-vated partial thromboplastin time (35.4 ± 1.5% after vs. 34.3 ± 1.1%before) were not reduced 1 h after intake of the mate infusion.These results were also confirmed in in vitro studies by incubatingplatelet-rich or platelet-poor plasmas with yerba mate extractprior to ADP-induced aggregation and coagulation tests. Additionof 800 lg/mL mate extract did not inhibit ADP-induced plateletaggregation or coagulation times (results not shown).

3.6. In vitro inhibition of copper-induced CE-OOH accumulation inplasma and LDL by mate extract and authentic phenols

The in vitro inhibition of copper ion-induced human plasma andLDL oxidation by yerba mate extract and its major phenolic con-stituent (chlorogenic acid), as well as a number of possible plasmametabolites of chlorogenic acid (caffeic, ferullic, and vanillic acids)was dependent on their concentrations, as shown in Fig. 4. Theinhibition of CE-OOH formation was calculated from the equation:Inhibition (%) = [(B � A)/B] � 100, where A and B are the concentra-tions of CE-OOH after 3 h incubation in the presence and absenceof either mate extract or antioxidants, respectively. The 50% inhibi-tion concentrations (IC50) of mate extract (as total phenol equiva-lents), chlorogenic, caffeic, ferulic, and vanillic acids for plasma

11 0 1000

20

40

60

80

100

Mate (eq.phenols) Caffeic acid Ferulic acid Chlorogenic acid Vanillic acid

Inhi

bitio

n of

pla

sma

lipid

pero

xida

tion

(% o

f con

trol)

Antioxidants (μg/mL)

Fig. 4. Inhibition of copper-catalyzed plasma lipid peroxidation by yerba mateextract and pure related antioxidants. Four-fold diluted plasma was oxidized with200 lmol/L CuCl2 at 37 �C for 3 h in the absence (control) or presence of mateextract, caffeic, chlorogenic, ferulic, or vanillic acids at the concentrations indicated.The data shown are the mean of duplicate tests and are representative of 3–4independent experiments for each variable. Standard error bars were omitted forclarity (variations were less than 6%).

978 E.L. da Silva et al. / Food Research International 41 (2008) 973–979

oxidation were 1.3, 6.1, 3.2, 3.5, and 10.4 lg/mL, respectively. Inthe copper ion-induced LDL oxidation, the IC50 values for mate ex-tract and chlorogenic acid were around 1 and 3 lg/mL, respec-tively. Similar IC50 values were found for mate extract andphenolic compounds in AAPH-induced plasma or LDL oxidation(data not shown).

4. Discussion

The main purpose of this study was to determine whether acuteconsumption of yerba mate infusion (I. paraguariensis) had inhibi-tory effects on ex vivo plasma and LDL oxidation and on plateletaggregation in humans. We found a lower susceptibility of plasmaand LDL to lipid peroxidation associated with a higher antioxidantcapacity after ingestion of a mate infusion. However, the observedeffects on ex vivo platelet aggregation and plasma coagulation werenot significant.

The remarkable ability of the I. paraguariensis extract to inhibitin vitro lipid peroxidation of plasma and isolated LDL has beenshown in several different oxidative systems including transitionmetal ions, free radical generators, and the lipoxygenase enzyme(Bracesco et al., 2003; Carini, Facino, Aldini, Calloni, & Colombo,1998; Filip, López, Giberti, Coussio, & Ferraro, 2001). In addition,it has been shown that I. paraguariensis is a more potent antioxi-dant than green and black tea or red wine (Bixby, Spieler, Menini,& Gugliucci, 2005; Bracesco et al., 2003; Campos et al., 1996). Thus,yerba mate might be considered one of the main antioxidant-richbeverages consumed in various South American countries, withan average ingestion of 1–2 L per day. About 3% of the weight ofI. paraguariensis leaves, or 1–10% of aqueous extract, are in theform of chlorogenic acid and its isomers (Bastos et al., 2007; Cariniet al., 1998; Filip et al., 2001; Heck & de Mejia, 2007), and the anti-oxidant activity of I. paraguariensis has been positively associatedto its chlorogenic acid and total phenol contents (Filip et al.,2000). The absorption and bioavailability of chlorogenic acid in hu-mans has not been fully clarified. Monteiro, Farah, Perrone, Trugo,and Donangelo (2007) identified chlorogenic acid and its isomersin human plasma after acute coffee consumption, a beverage richin chlorogenic acid, indicating their absorption and metabolism.In contrast, caffeic acid was the only phenolic acid found in humanplasma after coffee administration (Nardini, Cirillo, Natella, &

Scaccini, 2002). According to Olthof, Hollman, and Katan (2001),about one third of the chlorogenic acid ingested might be absorbedfrom the small intestine and the remaining reaches the colon,where the colonic bacteria probably hydrolyze chlorogenic acidto caffeic and quinic acids. Subsequently, caffeic acid is highly ab-sorbed in humans (Olthof et al., 2001) and metabolized to ferulic,vanillic and hippuric acids, which were found in urine (Rechner,Spencer, Kuhnle, Hahn, & Rice-Evans, 2001). Therefore, thesemetabolites of chlorogenic and/or caffeic acids might act as antiox-idants in the blood circulation. In fact, our in vitro results showedthat ferulic and vanillic acids possess antioxidant activity, althoughto a lesser extent than that of the parent molecules.

Overall, these data therefore corroborate the enhanced protec-tion of plasma against lipid peroxidation after yerba mate intake,which could be due to the presence of chlorogenic acid, caffeic acidand/or its metabolites. Moreover, regardless of the bioavailabilityof chlorogenic acid in humans, yerba mate aqueous extract con-tains vitamins C and E, and a number of other phenolic compounds,including flavonoids, which could play a role in the antioxidantcapacity of plasma. In fact, mate extract has been found to havea higher in vitro antioxidant activity than chlorogenic or caffeicacids, at the same phenol equivalent concentration, indicating apossible synergistic effect of mate constituents, as also reportedby Carini et al. (1998). Here, we showed the significant antioxida-tive capacity of yerba mate infusion intake in ex vivo oxidized plas-ma and LDL through more than one assay, measurement of CE-OOH and cholesterol oxides, which are highly implicated in theatherogenecity of LDL (Arca et al., 2007; Ferderbar et al., 2007).These results were obtained in association with a higher plasmaantioxidant capacity after mate consumption. Our results for plas-ma oxidizability are in agreement with the findings of Gugliucci(1996), who used other oxidation markers.

The present study showed for the first time that isolated LDLafter intake of mate infusion was more resistant to copper- orAAPH-induced oxidation, showing a slightly lower but still signifi-cant amount of the oxidation markers, CE-OOH and oxysterols. Inaddition to cholesteryl-ester, free cholesterol is one of the prefer-ential targets of oxidation in LDL (Arca et al., 2007) and the inhibi-tion of oxysterols formation by yerba mate infusion could be takeninto account as a potential mechanism to attenuate atherosclero-sis. The phenolic compounds of mate present in LDL preparationsmay contribute to inhibiting lipid peroxidation. In this regard,other authors have reported that phenolic compounds, includingchlorogenic and ferulic acids, might bind to proteins (Prigentet al., 2003; Harshadrai, Rawel, & Kroll, 2005). However, the spe-cific binding of mate phenols and/or their metabolites to apolipo-protein B-100 of LDL should be addressed in further studies.Finally, the acute ingestion of mate infusion had no effects on theinhibition of platelet aggregation and blood coagulation, also con-firmed by in vitro assays. As has been previously reported, humanplatelets in the circulating blood may play a role in the develop-ment of atherosclerosis, and increased platelet aggregation is asso-ciated with enhanced atherogenecity (Aviram, 1992). Therefore,our results indicated that yerba mate has no ability to attenuateplatelet aggregation and, consequently, might not play a role inthis atherosclerosis mechanism.

5. Conclusions

The results of this study showed that phenolic compounds fromyerba mate infusion were absorbed and promoted plasma and LDLprotection against ex vivo lipid peroxidation. Additionally, acuteconsumption of mate infusion had a mild but significant effecton antioxidant activity of plasma. However, ingestion of the mateinfusion did not decrease platelet aggregation or plasma coagula-

E.L. da Silva et al. / Food Research International 41 (2008) 973–979 979

tion time. Overall, the results showed that yerba mate intake mayenhance the acute antioxidant defense in humans, but long-termstudies are needed to demonstrate a more consistent in vivo anti-oxidant effect of I. paraguariensis and its relationship with inhibi-tion of atherogenesis.

Acknowledgments

This study was supported by a research grant (03/02832-1)from FAPESP (Fundação de Amparo à Pesquisa do Estado de SãoPaulo, Brazil). ELS was the recipient of a Post-Doctoral Scholarshipfrom the National Council for Scientific and Technological Develop-ment (CNPq, Brazil). We thank Dr. Daniel Falkenberg (BotanicalDepartment, Federal University of Santa Catariana, Brazil) for iden-tifying the plant material and Dr. Alejandro Gugliucci for criticalreading of the manuscript.

References

Arca, M., Natoli, S., Micheletta, F., Riggi, S., Di Angelantonio, E., Montali, A., et al.(2007). Increased plasma levels of oxysterols, in vivo markers of oxidativestress, in patients with familial combined hyperlipidemia: Reduction duringatorvastatin and fenofibrate therapy. Free Radical Biology and Medicine, 42,698–705.

Aviram, M. (1992). Platelets and arterial wall lesion. Current Opinion in Lipidology, 3,344–348.

Bastos, D. H. M., Oliveira, D. M., Matsumoto, R. L. T., Carvalho, P. O., & Ribeiro, M. L.(2007). Yerba mate: Pharmacological properties, research and biotechnology.Medicinal and Aromatic Plant Science and Biotechnology, 1, 37–46.

Benzie, I. F. F., & Strain, J. J. (1996). The ferric reducing ability o plasma (FRAP) as ameasure of ‘‘antioxidant power”: The FRAP assay. Analytical Biochemistry, 239,70–76.

Bixby, M., Spieler, L., Menini, T., & Gugliucci, A. (2005). Ilex paraguariensis extractsare potent inhibitors of nitrosative stress: A comparative study with green teaand wines using a protein nitration model and mammalian cell cytotoxicity. LifeScience, 77, 345–358.

Bracesco, N., Dell, M., Rocha, A., Behtash, S., Menini, T., Gugliucci, A., et al. (2003).Antioxidant activity of a botanical extract preparation of Ilex paraguariensis:Prevention of DNA double-strand breaks in Saccharomyces cerevisiae and humanlow-density lipoprotein. Journal of Alternative and Complementary Medicine, 9,379–387.

Bradford, M. M. (1976). A Rapid and sensitive method for the quantitation ofmicrogram quantities of protein utilizing the principle of protein-dye binding.Analytical Biochemistry, 72, 248–254.

Campos, A. M., Escobar, J., & Lissi, E. A. (1996). The total reactive antioxidantpotential (TRAP) and total antioxidant reactivity (TAR) of Ilex paraguariensisextracts and red wine. Journal of Brazilian Chemistry Society, 7, 43–49.

Carini, M., Facino, R. M., Aldini, G., Calloni, M., & Colombo, L. (1998).Characterization of phenolic antioxidants from maté (Ilex paraguariensis) byliquid chromatography/mass spectrometry and liquid chromatography/tandemmass spectrometry. Rapid Communication and Mass Spectrometry, 12,1813–1819.

Felippi, R., Wilhelm-Filho, D., Ribeiro-Do-Valle, R. M., & Silva, E. L. (2006).Administration of aqueous extract of Ilex paraguariensis reverses endothelialdysfunction in LDL receptor knockout mice. Free Radical Research, 40, S104.

Ferderbar, S., Pereira, E. C., Apolinario, E., Bertolami, M. C., Faludi, A. A., Monte, O.,et al. (2007). Cholesterol oxides as biomarkers of oxidative stress in type 1 andtype 2 diabetes mellitus. Diabetes/Metabolism Research and Review, 23, 35–42.

Filip, R., López, P., Giberti, G., Coussio, J., & Ferraro, G. (2001). Phenolic compounds inseven South American Ilex species. Fitoterapia, 72, 774–778.

Filip, R., Lotito, S. B., Ferraro, G., & Fraga, C. G. (2000). Antioxidant activity of Ilexparaguariensis and related species. Nutrition Research, 20, 1437–1446.

Geleijnse, J. M., Launer, I. J., van der Kuip, D. A. M., Hotiman, A., & Witteman, J. C. M.(2002). Inverse association of tea and flavonoid intakes with incidentmyocardial infarction: The Rotterdam Study. American Journal of ClinicalNutrition, 75, 880–886.

Giuliano, D. (2000). Dietary antioxidants for cardiovascular prevention. Nutrition,Metabolism and Cardiovascular Diseases, 10, 38–44.

Gugliucci, A. (1996). Antioxidant effects of Ilex paraguariensis: Induction ofdecreased oxidability of human LDL in vivo. Biochemical and BiophysicalResearch Communications, 224, 338–344.

Gugliucci, A., & Stahl, A. J. C. (1995). Low-density lipoprotein oxidation is inhibitedby extracts of Ilex paraguariensis. Biochemistry and Molecular BiologyInternational, 35, 47–56.

Harshadrai, M., Rawel, K. M., & Kroll, J. (2005). Binding of selected phenoliccompounds to proteins. Journal of Agricultural and Food Chemistry, 53,4228–4235.

Heck, C. I., & de Mejia, E. G. (2007). Yerba mate tea (Ilex paraguariensis): Acomprehensive review on chemistry, health implications, and technologicalconsiderations. Journal of Food Science, 72, 138–151.

Huxley, R. R., & Neil, H. A. W. (2003). The relation between dietary flavonol intakeand coronary heart disease mortality: A meta-analysis of prospective cohortstudies. European Journal of Clinical Nutrition, 57, 904–908.

Menini, T., Heck, C., de Mejia, E., & Gugliucci, A. (2007). Protective action of Ilexparaguariensis extract against free radical inactivation of paraoxonase-1 inhigh-density lipoprotein. Planta Medica, 73, 1141–1147.

Monteiro, M., Farah, A., Perrone, D., Trugo, L. C., & Donangelo, C. (2007). Chlorogenicacid compounds from coffee are differentially absorbed and metabolized inhumans. Journal of Nutrition, 137, 2196–2201.

Mosimann, A. L. P., Wilhelm-Filho, D., & Silva, E. L. (2006). Aqueous extract of Ilexparaguariensis (mate) attenuates the progression of atherosclerosis incholesterol-fed rabbits. Biofactors, 26, 59–70.

Muccillo-Baisch, A. L., Johnston, K. B., & Paganini-Stein, F. L. (1998). Endothelium-dependent vasorelaxing activity of aqueous extracts of Ilex paraguariensis onmesenteric arterial bed of rats. Journal of Ethnopharmacology, 60, 133–139.

Nardini, M., Cirillo, E., Natella, F., & Scaccini, C. (2002). Absorption of phenolic acidsin humans after coffee consumption. Journal of Agricultural and Food Chemistry,50, 5735–5741.

Neiva, T. J. C., Morais, L., Polack, M., Simões, C. M. O., & D’amico, E. A. (1999). Effectsof catechins on human blood platelet aggregation and lipid peroxidation.Phytotherapy Research, 13, 597–600.

Olthof, M. R., Hollman, P. C. H., & Katan, M. B. (2001). Chlorogenic acid and caffeicacid are absorbed in humans. Journal of Nutrition, 131, 66–71.

Prigent, S. V., Gruppen, H., Visser, A. J., van Koningsveld, G. A., de Jong, G. A., &Voragen, A. G. (2003). Effects of noncovalent interactions with 5-O-caffeoylquinic acid (chlorogenic acid) on the heat denaturation and solubilityof globular proteins. Journal of Agricultural and Food Chemistry, 51, 5088–5095.

Rechner, A. R., Spencer, J. P., Kuhnle, G., Hahn, U., & Rice-Evans, C. A. (2001). Novelbiomarkers of the metabolism of caffeic acid derivatives in vivo. Free RadicalBiology and Medicine, 30, 1213–1222.

Rimm, E. B., & Stampfer, M. J. (1997). The role of antioxidants in preventivecardiology. Current Opinion in Cardiology, 12, 188–194.

Silva, E. L., Tsushida, T., & Terao, J. (1998). Inhibition of mammalian 15-lipoxygenase-dependent lipid peroxidation in low density lipoprotein byquercetin and quercetin-monoglucosides. Archives Biochemistry and Biophysics,349, 313–320.

Singleton, V. L., Orthofer, R., & Lamuela-Raventos, R. M. (1999). Analysis of totalphenols and other oxidation substrates and antioxidants by means of Folin–Ciocalteu reagent. Methods in Enzymology, 299, 152–178.

Sokal, R. R., & Rohlf, F. J. (1995). Biometry: The principles and practice of statistics inbiological research (3rd ed.). New York: W.H. Freeman.

Stocker, R., & Keaney, J. F. Jr., (2004). Role of oxidative modifications inatherosclerosis. Physiological Reviews, 84, 1381–1478.

Triplett, D. A., Harms, C. S., & Koepke, J. A. (1978). The effect of heparin on theactivated partial thromboplastin time. American Journal of Clinical Pathology, 70,556–559.